OorsigOverview
Hierdie kursus is 'n inleiding tot digitale beeldverwerking. Die tegnieke wat behandel sal word is: beeldverbetering (beide in ruimtelike gebied en in frekwensie gebied), beeld-restourasie, kleurmodelle, beeldkompressie, en morfologiese beeldfilters. This course is an introduction to digital image processing. The techniques that will be covered are: image enhancement (both in the spacial and in the frequency domain), image restoration, colour models, image compaction and morphological image filters.
EvalueringEvaluation

Studente word in hierdie kursus word deurgaans ge-evalueer.

Twee toetse, elk rondom een kwartaal se werk, sal die wiskundige onderbou van beeldverwerking toets. Die toetse dra in totaal 25% by tot die finale prestasiepunt.

Ná elke onderafdeling word 'n projek uitgereik. Projekte dra in totaal 75% by tot die finale prestasiepunt. Projekte wat rekenaarwerk behels mag in enige oopbron sagtewarepakket voltooi word (Python, Octave, FreeMat, SciLab, C++, Mono, Java, ens.).

In this course, students are evaluated on a continual basis.

Two tests, each covering the work treated in a single term, test the mathematical underpinnings of image processing. The tests contribute a total of 25% to the final mark.

Following the treatment of each topic, a project is assigned. Projects contribute a total of 75% to the final mark. Projects that involve computer programming may be completed using any open source software package (Python, Octave, FreeMat, SciLab, C++, Mono, Java, etc.).

Test II Info

The final test takes place on Friday, 29 October at 10. This test covers all work done since the previous test, with emphasis on:

Dosente / Lecturers

Dosent / LecturerKantoor / OfficeE-pos / Email
Stéfan van der WaltA314stefan at sun
Willie BrinkA312wbrink at sun

Lectures

DateTopic
Third term lectures
14/09Reconstruction using a minimum mean-squared-error (Wiener) filter. The failure of norms in high dimensions. Overview of code for next assignment.
16/09Chromaticity diagram. Color spaces (CIE XYZ, RGB, HSV). Acquisition of colour images: Russia in colour. Manipulation of saturation: The colour of the moon.
17/09Manipulation of images in the different colour spaces. Colour slicing, histogram equalisation, segmentation, etc. (Read up to the end of Chapter 6).
21/09Introduction to morphology.
23/09No lecture. Work on assignment 4 and read up to the end of chapter 7.
24/09Heritage Day.
28/09Overview of segmentation. Point, line and edge detection. The Hough transform.
AwayWillie talks about image compression.
11/10Segmentation
15/10Boundary descriptors. Principle Component Analysis. Example: 33 bits to identify a person, Netflix de-anonymisation.
22/10Guest lecture: Ian Mason. Image processing in coal drilling.

Assignments

NrTopicHand-in
1Bilinear interpolation, geometric warping, homography estimation | Images: Chapel (grey)11/08
2Intensity transformations, histogram equalisation, smoothing filters, Difference of Gaussians, bilateral filter. Test images: Moon | Soil20/08
3Frequency domain operations Test images: Shifted Lena | Moon landing 01/09
4Linear, time-invariant filters Image: Motion-blurred clock 24/09
5Morphology, Hough transform11/10
General test images: F16 (grey), Tank (grey), Lena (grey)

Handboek / Textbook


Digital Image Processing
, Gonzales & Woods. Available from Protea & Van Schaik book shops. The first two chapters are available for download free of charge.

Lesings / Lectures (A308)

Tue 10:00-10:50
Wed 12:00-12:50
Fri 10:00-10:50

Buiteskakels / External links

Related Courses

MIT HST-582j Biomedical SIP